580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How can the probability and risk of Service Level Agreement (SLA) violations, especially regarding service availability, be modeled and quantified?
Quantifying the risk and probability of SLA violations related to service availability is crucial for providers to assess financial and operational risks. Research focuses on mathematical modeling of service states (availability/unavailability), statistical distributions of restoration times, and impacts of evaluation intervals on SLA breach probabilities and compensation liabilities.
Key finding: This paper develops a two-state availability model where service alternates between ON (available) and OFF (unavailable) states, assuming exponential distribution for ON durations and exploring exponential, Weibull, and... Read more
Key finding: Proposes a quantitative metric for measuring SLA fulfillment degree rather than a binary fulfilled/violated status. By analyzing SLA structural elements and monitoring data, the approach allows identifying critical service... Read more
Key finding: Defines foundational concepts and technical requirements for SLA and SLA management, emphasizing the importance of precise, measurable SLA parameters such as network availability, throughput, and latency to enable automated... Read more
2. How can multi-objective optimization methods improve SLA compliance and resource management in cloud and fog computing environments?
Cloud and fog computing face challenges in resource allocation balancing cost, energy consumption, and SLA compliance. Multi-objective optimization techniques, especially heuristic and meta-heuristic approaches, are investigated to optimize VM placement, server consolidation, and resource provisioning, simultaneously addressing conflicting objectives such as minimizing SLA violations, energy use, migration overhead, and operational costs.
Key finding: Introduces a multi-objective combinatorial optimization approach employing Non-dominated Sorting Genetic Algorithm II (NSGA-II) to generate Pareto-optimal SLAs balancing cost and makespan in cloud resource allocation. This... Read more
Key finding: Develops an enhanced bio-inspired locust swarm algorithm addressing VM consolidation and mapping in cloud data centers to reduce SLA violations, energy consumption, and VM migrations. Experimentation with real workload traces... Read more
Key finding: Proposes an AI-driven resource management policy using reinforcement learning to continuously consolidate and dynamically allocate VMs in cloud data centers, aiming to reduce energy consumption while maintaining SLA... Read more
Key finding: Presents an SLA specification framework optimized for fog storage services in IoT applications, focusing on a multi-objective trade-off among data availability, throughput, and cost. By specifying SLAs that incorporate these... Read more
3. What methods and protocols can enable dynamic SLA management including renegotiation and secure, automated SLA enforcement in distributed computing environments?
Static SLAs can limit flexibility and adaptability in dynamic cloud and grid computing environments. Research focuses on extending existing SLA specifications to enable runtime SLA renegotiation, dynamic adaptability, and automated secure enforcement. Blockchain-based consensus mechanisms and extended protocol specifications are explored to support trustworthy, decentralized SLA lifecycle management, including monitoring, violation detection, penalty enforcement, and renegotiation.
Key finding: Extends the WS-Agreement XML-based SLA specification to support dynamic SLA renegotiation at runtime. The paper introduces changes to the agreement structure, lifecycle operations, and states enabling SLA terms' adaptability.... Read more
Key finding: Proposes Proof-of-Monitoring (PoM), a blockchain consensus mechanism tailored for secure SLA (SSLA) management in multi-stakeholder 5G networks. PoM integrates SLA monitoring as a consensus task, reducing energy consumption... Read more
Key finding: Analyzes heterogeneity and semantic mismatches in existing service contract specifications, proposing guidelines to reconcile diverse contractual concerns spanning QoS, business, legal, and intellectual property domains. The... Read more